Chemical degradation of cementitious materials is a serious threat to the durability and performance of concrete structures. External sulfate attack is one of the situations that may cause gradual but severe damage. Sulfate ions present in seawater, rivers, groundwater and industrial effluent can penetrate into the hardened concrete, and react with cement hydration products to form ettringite as well as gypsum crystals, if stronger sulfate concentrations are available. Such formations result in a solid volume increase and cause local expansive pressure within the pore network. Although the solid volume increase may initially reduce the porosity of cement paste, it will cause cracking at a later stage as the generated expansive pressure exceeds the tensile strength of cement paste. This, in turn, leads eventually to a total strength loss and an increased permeability of concrete.
External sulfate attack in a saturated situation is a complex issue in which ionic transport, expansive reactions and mechanical damage interact with each other. These phenomena may be accompanied by significant macroscopic expansion and severe mechanical damage. However, the theories concerning the exact origin of the expansive pressure are still under debate. In recent years, the crystallization pressure theory has become the most widely cited hypothesis, and ettringite formation from monosulfate is also generally considered as the major cause, but more evidences for this mechanism are still needed. Moreover, the magnitude of the expansive pressure at different scale is still missing, since direct measurement of the expansive pressure on the walls of nanopores is highly challenging. Also the expansion behaviors of larger scale specimens are lack of complete experimental data fromthe literature. Furthermore, the process of crack initiation and propagation is seldom discussed. The development of pressure gradient has been largely neglected in the current literature.
In this thesis, an attempt is made to increase the body of knowledge related to cement paste expansion and degradation due to external sulfate attack. Laboratory experiments and numerical simulations were used during the study.
External sulfate attack under continuous immersion condition is a slow diffusion process. Even though high water/cement ratios and high sulfate ion concentrations have been adopted as acceleration methods, research shows that the attack depth remains shallow even after several months. Therefore, specimens with a small thickness along the diffusion direction could be preferred for experimental research in order to ensure a faster exposure of the entire cross-section. In this study, small cement paste pipes with a wall thickness of 2.5 mm were prepared and immersed in sodium sulfate solutions with SO42− ion concentrations of 1.5 g/L and 30 g/L. Three types of longitudinal restraints were applied on the specimens before exposure, which were created by a spring, a thin or a thicker stainless steel bar that was centered in the hollow specimens in order to facilitate the non-, low- or high-restraint condition. Strain gauges were used for the measurements of restrained expansions and generated stresses, with the purposes of increasing the measurement accuracy and obtaining continuous experimental results. The free expansion until 420-day immersion was measured periodically. The restrained expansion and corresponding generated stress until 810-day immersion were quantified continuously.
The pore size distribution, sulfur distribution and crack pattern were also periodically analyzed. According to the MIP measurements, the pores with diameters between 10 nm and 70 nm were continuously filled during the immersion tests, and strong sulfate solution lead to a faster filling, which supports the crystallization pressure theory. The sulfur distributions at 0-day, 21-day, 70-day, 105-day, 133-day and 189-day immersion in strong and weak sulfate solutions were acquired based on SEM-EDS microanalysis, which can reflect the gradient of the local expansive pressure distribution at the corresponding immersion days. The corresponding expansions under three types of restraint were also obtained. The specimen immersed in strong sulfate solution under high-restraint condition (7 mm - 30 g/L) was almost damaged after 565-day exposure with the largest generated stress of 13.4 MPa. Several vertical cracks were found after 581-day immersion based on the image analyses of CT scanning results. The specimen immersed in strong sulfate solution under low-restraint condition (3 mm - 30 g/L) was almost damaged after 628-day exposure with the largest generated stress of 11.2 MPa. One main vertical crack was found after 765-day immersion. The crack development of the unrestrained specimen immersed in strong sulfate solution (30 g/L) was also studied by CT scanning at 189-day, 294-day, 343-day, 420-day and 469-day exposure. A combination of the horizontal cracks which started some distance away from the exposed surface and the vertical cracks which started from the exposed surface was observed. However, no visually noticeable cracks were observed for the specimens immersed in weak sulfate solution (1.5 g/L) up to 807-day immersion. The generated stresses of specimens under low-restraint condition (3 mm - 1.5 g/L) and high-restraint condition (7 mm - 1.5 g/L) after 807-day immersion were 8.8 MPa and 11.1 MPa, respectively.
The complex process of crack initiation and propagation during material degradation at microscopic scale was studied by SEM - EDS microanalysis. The damage evolution of unrestrained specimens immersed in strong sulfate solution (30 g/L) was investigated experimentally. The specimens before sulfate exposure and after 70-day, 105-day and 133-day immersion were studied. Image analysis was applied. The localization process of the subparallel cracks near the exposed surface at the depth of about 250 μm was studied. Progressive precipitation of gypsum crystals inside the localized cracks was observed. The change of sulfur gradient versus exposure time was analyzed. Based on that, the change of expansive pressure gradient versus exposure time was discussed.
Numerical models can be of use in understanding complex problems. Delft lattice model was used in this thesis. In order to obtain the realistic mechanical properties of lattice elements, the experimental and numerical studies on the mechanical properties of cement paste pipes just after 90-day curing was done prior to further simulations. Two types of specimens are subjected to uniaxial tensile loading, which are unnotched and single notched specimens. Two main results were obtained through experiments. The first one is the Young’s modulus and tensile strength of unnotched specimens. The second one is the complete stress-strain curves of single notched specimens. A 3D lattice model with a mesh resolution of 0.25 mm/voxel was constructed to simulate the two types of specimens subjected to uniaxial tensile loading. After fitting with the experimental results, the local mechanical properties of cement paste lattice elements were obtained. Afterwards, a numerical study on expansion and degradation processes of the specimen immersed in strong sulfate solution under high-restraint condition (7 mm - 30 g/L) was performed. After comparing with the experimental results in previous chapters, the magnitude of local expansive pressure caused by external sulfate attack was discussed.
The experimental setup and techniques (such as strain gauge measurement system, SEM-EDS analysis and X-ray computed tomography) employed in this thesis can be used in the same or similar way for further studies on external sulfate attack or some other degradation problems. Also, the experimental results presented in this research can be used for further numerical studies.